Insights
Research, analysis, and perspectives on cybersecurity, AI security, and financial sector protection.
AI Security
Promptware Kill Chain: how we structure attacks against AI systems
Prompt injection is only the initial access. Our seven-stage framework to test AI systems end-to-end and map real 2025–2026 incidents.
Read articleAI Security
OWASP LLM Top 10: a practical guide for security teams
The ten OWASP LLM categories: what each means in practice and how to test them with reproducible evidence.
Read articleAI Security
MITRE ATLAS vs. ATT&CK: mapping attacks against models
ATT&CK describes the adversary in the infrastructure; ATLAS what they do to the model. How we cross-reference both in AI pentest.
Read articleAI · LLM · NVIDIA
From DGX Spark to 8x B200: How I Prototyped Locally and Trained a 4B Mamba-2 Model for €118
A practical journey from local prototyping on NVIDIA DGX Spark to training a 4-billion parameter Mamba-2 language model using 8x B200 GPUs on cloud infrastructure — all for just €118.
Read on Medium